Thursday, April 08, 2010


 Proto-alphabetic Inscriptions from the Wadi Arabah
Antiguo Oriente 8 , 2010 (expanded version of this essay)

The copper mines at Timna (in the Wadi Arabah/Arava, in Israel) have yielded a number of West Semitic inscriptions, and here are some interesting and instructive examples. They assist us in solving the problem of the identification of Sh, Ss, and Q in the proto-alphabet.


In August 2009 Joseph Otto of Stonewatch reported an inscription, and published a photograph of it; he would not reveal its precise location, but added that the Egyptologist Stefan Jakob Wimmer was working on it. It looks very Egyptian, with two 'cartouches', one showing an eye (of Horus?), and the other displaying a double serpent, protecting the sun, which could say Shimsh ('sun') in Canaanite, or simply be the letter Sh.

In the right-hand set of symbols, the circle under the eye (and its two lines) might represent the sun-disc. Or it may be a human head, R.

The two wavy lines symbolize water, and could the Canaanite MM (Hebrew mayim).

Is that a seated human person, or a human head (R) with a dot above it? As a logogram R'Sh ('head') it could say 'top-class' or 'prime quality' (as I have argued in two instances on the Wadi el-Hol inscription); hence it says "excellent water".  On the other hand, if it reads M MR, it becomes "bitter water". The eye at the top could thus be a logogram, or a rebogram, standing for `ayin 'spring'; three examples of such usage can be invoked from the Sinai mining region.

The other set of symbols is puzzling: 3 horizontal parallel lines, 3 oblique lines, and a character resembling the Egyptian nfr symbol (the lower part depicts the heart), standing for goodness and beauty, and borrowed in the proto-alphabet for, and in the syllabary for T.A, or as a logogram t.ab 'good'. The trio of strokes might represent the spinal column (the Egyptian djed, 'stability'), used for the letter Samek ('support'), though the fish-symbol is more often found for Samek, until the Phoenician alphabet settled for the column S (and then into the Greek alphabet for Xsi).

The cartouches suggest that an important person  would be named in the text. But is this actually a 'sign' saying: "spring",  "excellent water", "good support"?

As always, we cannot be sure about the writer's intended meaning. But on Vladimir's photograph there are more marks: possibly N (snake) and G (throwstick), and others below them, not completely in the frame. If we can read GN (gan 'garden') then we would have a counterpart for the Garden of the Sun at the Sinai turquoise mines (Sinai 353), and this could explain the sun-symbol here.

If this hypothesis is correct, the site should have an ample patch of ground for a vegetable garden, and be near the camp of the miners. The Sinai horticultural inscriptions were on the exterior wall of a mine, indicating that the garden plot was right there. If this supposition fails, a new context for this inscription must be sought. The possibility remains that it is marking a well, but again the feasibility of this idea needs to be tested by an examination of the site.

Stefan Wimmer has now provided a full description of the place and the text, in JOURNAL OF ANCIENT EGYPTIAN INTERCONNECTIONS (online,| Vol. 2:2, 2010| 1–12 pdf).

 Stefan Wimmer's photographs show that there is ample space for an irrigated vegetable garden; and the inscription (its position is indicated by the white arrow on the left, and the seated man on the right, who is in a pose similar to that of the human figure in the inscription) is actually horizontal (not vertical, as other pictures of it may have suggested) and situated on a tabletop rock (about 2 metres high). Whether there was ever a spring on this spot is not a question I can answer.

Here is a copy of his drawing.

Stefan's interpretation is interesting and authoritative: the two ovals are not Egyptian cartouches enclosing royal names (they lack the mandatory horizontal line at the bottom); the writing is not Egyptian but belongs to the same family as the pictorial characters from the Sinai turquoise mines, the so-called 'Proto-Sinaitic' script, but  I would refer to it as  the West Semitic proto-alphabet or the Proto-Canaanite pictophonographic consonantary, represented by numerous inscriptions in Syria-Palestine, Egypt, and Sinai (including Timna and Har Karkom), from the Bronze Age.

The ovals could depict tablets (like the two tablets of Moses), and SJW proposes to find a word for 'tablet'  in the one on the left; or else footprints, with the name of the person engraved on them (and this might make it a 'Kilroy was here' graffito).

The oval on the right has the following signs: `(ayin), Z (properly Dh), R (head), M M, and the human figure is a classifier borrowed from the Egyptian repertory, indicating a male person.  So we would have a  man's name: ` Z R M M.

The other oval has the sun symbol at the top (Sh, from shimsh 'sun', hieroglyph N6B, two uraeus serpents encompassing the sun, though here the sun-disc is omitted), and Stefan has accepted my argument (first published in 1988 and reiterated and reinforced ever since, but almost totally ignored by scholars in this field) for seeing hieroglyphs representing the sun (r`) as the source of proto-alphabetic Sh-signs. He compares the character appearing twice on the Wadi el-Hol inscription (see my notes on Sh-signs there) with two circles joined by a curved line (which he has always thought to be Sh, but without making this connection), and he sees that as a variant of what we have here: the  two snakes without the sun disc. However, whereas here the serpents have heads of the same size, there the scribe has in each case made the circle on the right considerably smaller than the one on the left, suggesting that one represents the sun-disc and the other is the head of the snake. Thus, as I see it,  there we have an example related to hieroglyph N6, with only one serpent, but here we have N6B with two.  This is a crucial point for dating: the hieroglyph with two snakes (N6B) is perennial (known in the Middle Kingdom and the New Kingdom), but the single-serpent icon does not appear before the New Kingdom, and this means that the Wadi el-Hol inscription can not be assigned to the Middle Kingdom, where everyone wants it to be,  so that it can maintain its supposed position as the oldest alphabetic inscription. 

Now, the typical Sh-sign in the Sinai collection does not include the sun-disc, and because it curls round at each end, the two-serpent hypothesis seems to fit the case admirably, rather than seeing a head at one head and a curly tail at the other; N6 has a straight tail, and this does not appear on the Wadi el-Hol examples.

However, the disc is not always omitted, as shown on the inscribed Timna stone that I am examining here (see below): head, sun, and straight tail.

Or on this potsherd from the Valley of Queens in southern Egypt.

My reading for the two lines (right to left):
' M H T (maidservants), ' Sh T (women)
The lower line should be viewed vertically to see the ox-head (`aleph), the sun with a single serpent (Sh), and the cross (T).
(See Benjamin Sass, Genesis of the Alphabet, 1988, figure 286, and p.104, but he disallows this as proto-alphabetic, contra Leibovitch, Albright, van den Branden, and myself).

Then an example from Thebes (5th character from the left), which (although it is in a syllabic inscription, as ShI) seems to show how the form without the disc came into being.


(William Flinders Petrie, in the frontispiece of his book The Formation of the Alphabet, 1912 )

Of course, even though this Timna inscription could not be a hoax (see Stefan's arguments, p. 7-8), we might actually be looking at two eyes, a nose, and a mouth, and all this agonizing over the origin of Sh (which, incidentally, did not find its way into the Phoenician and Grecian alphabets) crumbles to dust. On the other hand, if it is correct, then all the speculative theorizing about Sh as a thorn and Sh/Th as a composite bow (following W. F. Albright, most recently in Gordon Hamilton's The Origins of the West Semitic Alphabet, 2006, 123-125, 231-244) collapses in ruins.

However, the three horizontal lines below the sun-symbol are not a mouth but a hand, in the simplified three-finger form often used for K; Stefan has observed (on the site) a line joining the three digits on the right, and this is included in his drawing.

Significantly, the failure to recognize the solar connections of the sign for Sh/S (cp Hebrew Shin/Sin) caused the exponents of the composite-bow hypothesis to read Th every time it occurred, producing Dh T B Th N 'the serpent woman' in Sinai inscription 351, instead of Dh K B Sh N M Sh 'this melt-furnace' (Hebrew kibshan 'furnace', with Shin; and mss 'melt', with Sin); the K has three fingers, pointing down. The same term reappears in other inscriptions, in a material context of metal-working equipment, and this is one of the keys to the decipherment of these documents.

Moving on, the sign on the bottom left  is understood as L, a crook (like Hieroglyph S38 rather than S39 which is simpler, not curved backward), though it is inverted here; it is not so likely to be in imitation of V1, the coil of rope favoured by Hamilton (126-137).

Finally, the three oblique strokes are provided with a stem, and instead of making a connection with the Egyptian djed column (R11) and the letter Samek (as I suggested above), or interpreting it as another K, he invokes the hieratic version of V28 (ooo<) with strokes not circles for this double-helix sign. This is an original idea and worthy of consideration. However, there is a serious mistake here: although this hieroglyph stands for Egyptian Hh (dotted h), it does not have this value in the proto-alphabet, but Kh (see Hamilton, 57-60). He seems to imply that this is the sign used for Hh in Sinai, but not in Canaan. However, this character (a hank of thread or a wick) represents Kh in proto-alphabetic inscriptions in Egypt, Sinai, and Canaan, and also in the cuneiform alphabet, where Kh is a cluster of three vertical wedges, obviously based on this pictophonogram (see the Kh and Hh lines on my table, at the end of this article).

The resulting sequence is stated as Sh K L Hh (properly Kh).
The proposed interpretation is S´K L (L) Hh 'tablet-expert' (cp Hebrew s´kl 'have insight'; luah. 'tablet')
The difficulty is that the Ugaritic evidence clearly  has LHh not LKh for 'tablet'. Nevertheless, there may be a way out of this dead-end:  in the shorter 'linear' version of the pictorial alphabet the sign for Kh disappeared, and Kh coalesced with Hh, but in the shorter cuneiform alphabet Kh replaced Hh.
This inscription is presumably from the  Ramesside period, when the Bronze Age is ending and the Iron Age is commencing, and things are in a fluid state in the evolution of the alphabet and the phonology of the Canaanite dialects.

This principle could be used to cover the Kh for Hh here in oval L, and the Dh for Z in oval R. The inscribed stone studied below has ShQL 'weight, shekel', but this should be ThQL in a Bronze-Age setting; but Th (breast) is the one that survives in the alphabet. This is a limited corpus to work with, but the Hh sign is lacking in it, and again this will be the one that remains, while the Kh (thread, wick) will disappear.

Now, it is a truism that only the writer of an ancient inscription knew what it meant, and I and Stefan have proposed two vastly differing interpretations. His was a typical autograph graffito: " `Az-romam the scribe". Mine was a notice concerning a spring in the immediate vicinity, and the state of its water.

SJW has achieved his reading by doubling the L, and for this he refers us to the sequence M'HB`L[T] on the Sinai sphinx statuette (S345), standing for M'HB B`LT 'beloved of Ba`alat'. But nobody ever mentions that the B (a square representing a house, bayt) has a dot in it, presumably indicating the double B (as in classical Hebrew writing).

 The L here possibly has such a dot beside it, but so does the K. Now, a Hebrew root ShKK appears in the Flood story in Genesis (8:1): 'the waters abated'. When we add to this the fact noted by SJW (5b) that three parallel straight lines (instead of wavy lines, as in oval R) is the hieratic form of mw 'water', we have the waters abating here also!

The first thoughts I had on this inscription might merit recording here. 

Allowing that the right side is telling us about a water-well (`ayin 'spring', M 'water'), the left side might contain a reference to an important piece of apparatus, namely a bucket for drawing water out of the well.  The Semitic root is DL (Akkadian dalu 'draw water', 'bucket'). The L is on the bottom left, and the two oblique lines beside it may have joining strokes, making a door, hence D (Dalet). The top line of this trio could be a simple snake, N, preceded by K and Sh. The combination ShKN produces the root that means 'put, place' or 'dwell' or 'be present'. The word MShKN (used for the 'tabernacle' of the Bible, the tent where the Shekinah 'presence' of God dwells) is found at the camp site of the Egyptian turquoise expeditions in Sinai (S365). Here it signals the place where the bucket is to be placed after use.

If there is no possibility of a well having been situated here in the remote past, then this line of interpretation must be rejected. Would that the author of this puzzle had left the solution under the rock.


Beno Rothenberg, The Egyptian Mining Temple at Timna (London 1988)
Plate 116 : 4 and 5
Sandstone pebble (6.5 x 7.5 cm) with engraved signs on both sides; unstratified (p. 268b)

This is possibly a weight, and indeed the word shekel (ShQL) appears on it, apparently.  The Q (-o-) is in the middle of the object, on the first side. It is surely not necessary for me to go into a long defence of this identification, even though the handbooks on early West Semitic writing have overlooked it, and assigned the value q to the tied-bag sign, which is actually S.adey. Because S.adey and Qoph are relatively rare letters and difficult to identify,  this school of thought  takes the Hebrew name Qoph, meaning 'monkey', and turns the bag into an ape. However, Q survived (in the form seen here) in the old Arabian script; and moved into the Phoenician consonantal alphabet without the top projecting stroke (see the Q line on the table at the end). It represents a line (qaw), a cord wound on a stick, and it has an alternative form, imitating its borrowed  Egyptian hieroglyph (V24, V25): an end of the cord pokes out at the top (see the table), and this is found on the sphinx (Sinai 345) from the Sinai temple of Hat-Hor (an earlier counterpart to the shrine at Timna); it has been unrecognized by previous observers (but note the dot in the middle of the main stem, and the other projrcting line on the left). This later form does not appear till the New Kingdom in Egypt (Late Bronze Age), and this should mean that the inscription on the sphinx can not be dated to the Middle Kingdom (Middle Bronze Age), as many have assumed. The example we see here seems to have only one projecting line at the top, though it is easy to imagine another one among all the marks. The development of the letter Q involves elimination of all lines above the circle, as in Roman Q. There is another example on the inscribed plaque, below (in an unusual horizontal stance).

The Sh sign is the sun with a serpent; the tail is on the right, and the head on the left (see the Sh/Th secton of the table, below). The presumed  L stands next to the Q; it is a simple crook (see the L line on the table).

There is a fish below the Sh, which would represent S (not D for dag 'fish', as commonly supposed; D is from dalt 'door', and there may be one on the other side of the stone). To the right is a right-angle, presumably a boomerang, and thus G (gaml, not P).

On the other face, three signs are detectable. From the left: a door, D; an ox-head, 'alep; a hook, W, or a head, R (but the neck is too long), or another L with the top more curled than the other.

If the sequence is S G D ' L, then it can be analysed as SGD 'worship' and 'L 'El' (God), though a preposition l ('to') would be expected. As a guess, are these extra words somehow giving divine sanction to the trustworthiness of the weight?


Stone plaque from the Arabah
Photograph and drawing as reproduced by Sass (fig 276, 277)
See Sass, The Genesis of the Alphabet, p. 103 for information
Assuming that this is another West Semitic inscription, and giving credence to the drawing: the bottom sign is T (+), beneath N (nahhash 'snake'), then Q (qaw, cord wound on a stick, as on the stone, above); the most likely identity for the next one is Ss (s.irar, a tied bag, see the Egyptian original on the table, below); the wavy line is a short M (water), or possibly Th (breast); at the top we have a sign shaped like H, which is a late example of Z (not Dh). There are other possible letters above and beside this Z.

There are more marks to the left of this column of writing. Level with the Q is possibly an eye (signaling another spring, as possibly in the rock inscription above?).

The sequence of signs on the drawing would be:  Z M Ss Q N T

In the light of the mining and smelting of copper that was done in this area, a connection could reasonably sought with the root ys.q (Y Ss Q, Ugaritic and Hebrew) 'pour' or 'cast' metal. If the MSsQ means "(place of) pouring" or "casting" (cp Hbr, then NT could be analysed as a rebogram plus a consonantogram: NHhSh (the consonants of the word for 'snake') + T, producing NHhShT, "copper" or "bronze". If Z is 'this' (though Dh would be expected in the Bronze Age, but these mines belong in the Ramesside period, when some consonants are coalescing) we could interpret the statement thus:
  Z MSsQ N(HhSh)T
 "This is the copper-smelting (place)"

Incidentally, I trace the Z-sign to ziqqu 'manacle', and as a rebogram it can represent the root zqq 'refine'; this term appears in connection with refining gold on an inscription from Thebes (ostracon 4).

The snake rebus for copper (possibly also on Sinai 352) would show that the characters of the proto-alphabet could be used like Egyptian hieroglyphs, that is, as logograms and rebograms (rebuses).


These three inscriptions seem to be giving support to ideas I have been promoting for many years, notably that the proto-alphabet was in a way a simplification of the Egyptian writing system.

They also bolster my identifications of the controversial signs: there are two instances of the true Q (qaw, 'line', string on stick), and (on the plaque) one of them is right next to the false Q (tied bag) which is actually Ss, and occurs in the Sinai texts in such words as s.btm 'handfuls' (375), s.rp 'crucible' (372), s.rh. 'excavation chamber' (356), and ns.b 'foreman' (346, 349, 351). The choice made by Albright and his followers to read nqb 'piercer', supposedly meaning 'miner', instead of ns.b 'prefect', sent them off on a wrong track, on which they were compelled to overlook the occurrences of the real Q, in ql` 'inscribe' (376), qnt 'elegy' (363), nqy 'my offering' (345, the sphinx). 

The confusion of the door as D (dalt) with the fish as S (dag 'fish' is invoked to justify it as D) can be refuted when both are found in the same text; there is a faint possibility of such an occurrence on the small stone, but a clear case is on the wall above Bir Nas.b, the well where water was obtained for the turquoise and copper expeditions, in the words DWT 'sickness' and 'S' (Asa) (Sinai 376).

One way of confirming that D is a door is by examining the rectangular shape it has in the cuneiform  alphabet (see the full cuneiform table, and also the Canaan column of the table): D has 3 vertical wedges (the sections of the door) on 3 horizontal wedges (the door post); B, the square house, has four wedges, similarly arranged. Compare also the form for S, which has 3 wedges, representing the fish, very simply, with head and tail. Note that the alternative form of Samek (spinal column), which possibly appears in the rock inscription, also has a place in the cuneiform alphabet.

Comparing the cuneiform  S. and Q, the Q has a horizontal wedge (note the -o- form on the plaque, above) with a type of wedge that usually represents a circle in the original pictogram (it is employed for the eye of `ayin); S. has two vertical wedges, representing the bag standing upright, presumably.  Additional confirmation comes from the Arabian forms: Q has retained the cord-and-stick, and S. is the bag.

Sh (sun) and Th (breast) both have the wedge representing roundness; but Sh has 3 parts, for sun and serpent head and tail, or two snake heads; the form with two sun-serpents (N6B), seen on the rock inscription above (but with the sun-disc omitted), seems a better model for the typical form of Sh in the Sinai inscriptions, I would now say. Nevertheless, on the stone above we see a form of N6A, with one head and a tail. The 2 examples on the vertical inscription of Wadi el-Hol (shown on the chart) both have a small head and a large sun, and no tail. This version resembles the Arabian Th (o-o), and it appears that Sh and Th have exchanged places in the Arabian script.

Click on this table for an extra-large version

A darker version may be viewed here


Michael Sheflin said...

Great comments so far. As far as I can tell, the first character on the left is the only paleographic analogy both to the character at Wadi el-Hol and in the Arabian corpuses; obviously in the former I read this as T_ and you as S^. However, I believe it is unmistakably S^ here. Additionally, the final character on the right of the left-most tablet may be paleographically identical to the Ugaritic "'i."

If these features are correct, then the D_ would have already been dropped both from script and language most likely, and the double-lines that are D_ in Proto-Sinaitic would thus be an archaic form of the Punic Z. It doesn't really matter, you could read the same thing with Z=D_ 3D_R is help in Ugaritic, and though the enclitic demonstrative is usually rendered D in Ugaritic, D_ is still possible.


S^ This year

L<-I Il



MMY Memmy

May Il aid Memmy this year.

Z4chst3r said...

I sent the following comments to Wimmer on regarding a very different reading to what he gave in his paper:

Some thoughts on "A Proto-Sinaitic Inscription in Timna/Israel" and a suggestion of a very different reading:

For the right inscription: There was a distinction in the early alphabet between z and dh. The two horizontal lines is dh not z so it is the wrong consonant for either `-z (strong) or `-z-r (help). Unless of course we are dealing with a case where in the dialect of the writer the two sounds had merged and he on used the dh sign. But going with the usual understanding as dh we see the root `-dh-r meaning plea. The double m is understandable as water = maim. The seated figure looks most like Egyptian determinative A7 denoting resting not a personal name - this would indicate that the inscription relates to resting.

For the left inscription: There were several different non-emphatic sibilants in central Semitic. The recurve bow T9A (others say its breasts) corresponds to the sibilant that became th in Arabic but sh in Hebrew and t in Aramaic. Then there was one that became sh in Hebrew but s in Arabic, this appears in Sinaitic as M44 the thorn which appears to be the forerunner of the Ethiopic letter "sat". Then there is the sibilant that became s' (sin) in Hebrew but sh in Arabic. At Wadi El-Hol I see this represented as the sign that appears to be a pair of lips (vertically oriented so perhaps labia rather than a mouth) - Wilson-Wright also identifies it as s' but attempts to explain it as a hieratic sun rather than lips but that would be the wrong sibilant (sephataim - lips makes more sense). There is also the sibilant (probably affricate originally) that became s (samekh) in Hebrew and also s in Arabic - that appears at Sinai as the fish but other early West Semitic inscriptions have the djed pillar for this. But there is also the sibilant that became/remained sh in both Hebrew and Arabic, the initial sibilant of shemesh/shams = sun. This is what the uraeus sun sign that starts the left inscription represents sh not s'. (It is not the same sign found at Wadi El-Hol with two loops - that as per Vanderhooft is instead the yoke / tether V13 ghol/ghullu for the gh sound, the South Arabian gh is clearly based on the hieratic form of V13). Also although the bottom right sign resembles somewhat the hieratic form of the the hank/wick V28 that is the wrong guttural for luach meaning tablet - luach would have the mansion 06 for the pharyngeal H while the wick is velar kh. Also here the orientation of the coil / coiled crook suggests it is to be read last if we compare other West Semitic inscriptions. It is more likely that rather an hieratic wick we simply have the djed pillar samekh for s. I thus read the left inscription as sh-k-s-l. How should this be parsed? The only sensible split is as two words sh-k followed by s-l. S-l has an immediately recognizable meaning as a basket or pannier - the term surviving into modern times. For sh-k here we need a root where the sibilant is the same as the initial sound of shemesh/shams - it should appears as sh in both Hebrew and Arabic, not as th or s in Arabic and only sh in Hebrew. We do have a such a root sh-k(-k) which in Hebrew meant to stoop or to incline a container - in Arabic the root survived with only an idiomatic meaning of to wait hesitantly whence it has acquired the meaning of to being doubtful of something.

The full inscription would then read "incline (or wait with) a pannier, make a plea for water" a sign post indicating an oasis where water would be offered to those who asked.

Z4chst3r said...

Another comment I have sent to Wimmer:

"Another thought on the root `-dh-r, difficulty in understanding it is that although there was a distinction between z and dh these seem to have originated via inter-dialect and inter-lingual borrowing so the meaning of plea is ultimately related to meaning of `-z-r of help. But one particular usage of the root is for a word meaning a container, so it is possible that the word here is meant to be paired with s-l with the expression s-l `-dh-r denoting a type of container, corresponding to ḫa-zu-ra in Akkadian, with the final word mm further denoting that it is for water. The whole inscription might then mean something like "tip water carry container"."